Add support for Unicode identifiers in GDScript

This is using an adapted version of UAX#31 to not rely on the ICU
database (which isn't available in builds without TextServerAdvanced).
It allows most characters used in diverse scripts but not everything.

modules/gdscript/tests/scripts/parser/errors/identifier_similar_to_keyword.gd

@@ -0,0 +1,3 @@
+func test():
+	var аs # Using Cyrillic "а".
+	print(аs)

modules/gdscript/tests/scripts/parser/errors/identifier_similar_to_keyword.out

@@ -0,0 +1,2 @@
+GDTEST_PARSER_ERROR
+Identifier "аs" is visually similar to the GDScript keyword "as" and thus not allowed.

modules/gdscript/tests/scripts/parser/features/unicode_identifiers.gd

@@ -0,0 +1,35 @@
+const π = PI
+var ㄥ = π
+
+func test():
+	var փորձարկում = "test"
+	prints("փորձարկում", փորձարկում)
+	var امتحان = "test"
+	prints("امتحان", امتحان)
+	var পরীক্ষা = "test"
+	prints("পরীক্ষা", পরীক্ষা)
+	var тест = "test"
+	prints("тест", тест)
+	var जाँच = "test"
+	prints("जाँच", जाँच)
+	var 기준 = "test"
+	prints("기준", 기준)
+	var 测试 = "test"
+	prints("测试", 测试)
+	var テスト = "test"
+	prints("テスト", テスト)
+	var 試験 = "test"
+	prints("試験", 試験)
+	var പരീക്ഷ = "test"
+	prints("പരീക്ഷ", പരീക്ഷ)
+	var ทดสอบ = "test"
+	prints("ทดสอบ", ทดสอบ)
+	var δοκιμή = "test"
+	prints("δοκιμή", δοκιμή)
+
+	const d = 1.1
+	_process(d)
+	print(is_equal_approx(ㄥ, PI + (d * PI)))
+
+func _process(Δ: float) -> void:
+	ㄥ += Δ * π

modules/gdscript/tests/scripts/parser/features/unicode_identifiers.out

@@ -0,0 +1,14 @@
+GDTEST_OK
+փորձարկում test
+امتحان test
+পরীক্ষা test
+тест test
+जाँच test
+기준 test
+测试 test
+テスト test
+試験 test
+പരീക്ഷ test
+ทดสอบ test
+δοκιμή test
+true

modules/gdscript/tests/scripts/parser/warnings/confusable_identifier.gd

@@ -0,0 +1,5 @@
+func test():
+	var port = 0 # Only latin characters.
+	var pοrt = 1 # The "ο" is Greek omicron.
+
+	prints(port, pοrt)

modules/gdscript/tests/scripts/parser/warnings/confusable_identifier.out

@@ -0,0 +1,6 @@
+GDTEST_OK
+>> WARNING
+>> Line: 3
+>> CONFUSABLE_IDENTIFIER
+>> The identifier "pοrt" has misleading characters and might be confused with something else.
+0 1